Get reliable science

If the landscape function paradigm is to fulfil expectations, then it must, by definition, make ecological sense at the landscape -scale, as distinct from the patch -scale, and thereby provide meaningful direction to land managers. The approaches so far suggested for interpretation of an ecosystem's functionality do not have universal application. Our overall impression is that the work on developing indicators of landscape function has moved well ahead of the underpinning science. We must do the basic research work necessary to modify our rangeland models so they pass the 'test of predictive power'. We require a good understanding of fluxes of water, soil particles and associated nutrients and litter within and throughlandscapes and  how these fluxes vary in response to, for example, geomorphology, vegetation structure and soil type. Just as we are learning that landscape function cannot be generalised across all landscapes, so we may learn that the interactions of composition, structure and function cannot be generalised to support the management of biodiversity. Our interpretation of the papers presented here is that our ability to interpret impacts of management of rangeland and thereby prescribe alternatives remains elusive.